Showing posts with label Collaborative Work. Show all posts
Showing posts with label Collaborative Work. Show all posts

Sunday, December 23, 2018

Process Alphabet: Information Rolling Wave

Draft Version 2

 

Intro

Our subject of work is the information
Information rolling wave is the model of what we do

Life-cycle, practices, context, delivering value - all must fit together in our information rolling wave. Our subject of work is the information that is gathered, analyzed, produced, validated and used. Looking ahead and feedback dimensions should be both represented for all disciplines and all timing levels by appropriate practices   

Expected support from a method


·        Information states
  • Motivation - our primary work object is information and progress should be associated with specific states for “crafted” information
  • Subjects of states - requirements, design, implementation, working software, solution & its parts, etc
  • Types of states - conceptual, feasible, ready, done, consumable, production-ready ...
  • Waterfall versus Iterative/Agile
    • Waterfall consider states per each discipline (e.g state of requirements)
    • Iterative & Agile consider the consolidation of the states at the work level. Work items should be based mostly on working software and the progress should relative only the realization of Consumable Solution.
·        Adapt to Context
  • Roadmaps & Life-cycles - Motivation: we need a high-level evolution shape for the information that we gather and produce
  • Process Goals - Motivation: across the life-cycle, we shape the information wave around process goals
  • Selecting practices - Motivation: we need to make the optimum decision in context ~ selecting the right practice. For selection, we need a library of practice and guidance about the tradeoffs of using different practices. Practices will shape our information rolling wave. TDD example: offers a direct trace from requirements, test cases and realization, an explicit & quick validation; by adding/removing TDD, our information rolling wave could change dramatically: from one-month potential release to one day or less.    
·        Collaboration
  • Motivation: Knowledge work has collaboration as a core component: people are the most important “repository”, “processor” and “channel” for information management. Collaborative work dramatically changes and improve the information rolling wave
  • Types of collaboration: Inside team, between development teams, between development teams & other teams
Value stream shaping the rolling wave 
  • Motivation: Value-stream mapping - “Material and Information Flow Mapping” at Toyota - in the case of software development is mainly information flow mapping. Value stream helps us to identify the useful types of work and subsequently the needed practices
  • Core practices (see below) are proven domain-level instruments that fit the development value stream; for each particular situation we still need to adapt practices, life-cycles & to context      
  • Value stream support examples: pragmatic product-oriented practices or better consumable solution support. Focusing on the final output -  the product - helps to avoid unnecessary work. Focusing on the final outcome - a consumable solution is even better.   
  • Clear support for information states
·        Core Practices shaping the rolling wave
  • Motivation: we need to give special care to practices that shape the information rolling wave, the ones related to Looking Ahead and to Validation & Feedback. We can dramatically optimize the work if we can use the Core Practices related to these aspects.
  • Look-Ahead practices: envisioning, look ahead, just-in-time
  • Validation and feedback practices: from reviews to automated tests
 
Rolling Wave in Scrum

·        Information states
  • Explicit state for work item at iteration level only: ready, done
  • Vague guidance for ready and done
  • Potentially Shippable concept does not explain well all aspects of a Consumable Solution
  • No states at the life-cycle level    
·        Roadmaps & life-cycles
  • See Process Alphabet: Life-cycles
  • Roadmaps - Product Backlog can be used as a development roadmap, but no other explicit support for roadmaps, life-cycle, and life-cycle types
·        Adapt to context
·        Collaboration between development teams
  • Scrum basic has nothing
  • Scrum core-extensions (from Scrum creators) offer only Scrum of Scrums option
·        Collaboration with other teams
  • Very short guidance about direct customer collaboration
  • Nothing about collaboration with other enterprise teams
Value stream support 
  • Product backlog, product increments oriented practices (the spring main goals is to produce value via product increment)
  • What is missing
    • alternative variants for different situations (e.g missing parts on Life-Cycles)  
    • focusing on the outcome (Consumable Solution) versus focusing on the output (working software) 
·        Selecting practices that shape the rolling wave
o   Prescribed practices, without options
o   Only iteration, iteration meetings, daily meetings, product backlog (and items)
o   Looking ahead – vague description as “refining”. 
o   Too few guidance for envisioning and no guidance for just-in-time practices
·        Selecting practices for continuous validation and feedback
o   Prescribed practices, without options
o   Too few: iteration-level only + daily meeting. No just-in-time practices and no info about automatic tests, continuous integration, continuous deployment

Scrum advantages
  • Scrum allows adding other practices when it does not displace its prescribed practices. Anyway, Scrum does not provide guidance for that
What is missing
  • You cannot replace Scrum prescribed practices (and still doing Scrum)
  • Continuous Delivery and Lean life-cycles with very small iterations are not supported by Scrum. Information rolling wave and validation ceremonial in Scrum has a much higher granularity, starting at the week level. Comment: you can modify a Scrum-based process to support these life-cycles, but the result is no longer Scrum.
  • No support and guidance for low granularity validation and feedback that contribute to the effectiveness of “Done”
  • Progress toward a Consumable Solution has not enough support
  • Very few guidance in general and no guidance about full, end-to-end life-cycle rolling wave model  


Rolling Wave in XP


·        Information states
  • Explicit state for the work item: ready and done (working software, integrated and tested by default)
  • Work envisioning at release/iteration level in release/iteration planning
  •  State of the product: exploration, first release, maintenance
  • Progress toward a Consumable Solution is not so explicit, excepting the working software part 
  •  Feasibility & pragmatism of the states
    • Short releases with short and often internal cycles
    • Product built-in quality
    • Simple Design avoids complexity & irrelevance of states
·        Roadmaps & life-cycles
  • See Process Alphabet: Life-cycles
  • Roadmaps: short guidance about product life-cycle: exploration, first release, maintenance and short guidance about quarterly planning 
·        Adapt to context
·        Collaboration between development teams
  • Very short guidance: “integrate frequently” and planning could help
·        Collaboration with other teams
  • Guidance about direct customer collaboration: On-Site-Customer and Real-Customer Involvement.  Nothing about collaboration with other enterprise teams  
Value stream support  
  • Small releases with often adaptation on the changing business needs
  • Implicit support by strongly use of the core practices
·        Selecting practices that shape the rolling wave
  • XP has a significant part of this kind of core practices: small releases, envisioning, iterations, weeks management, daily meetings, pair programming, TDD, continuous integration
·        Selecting practices for continuous validation and feedback
  • XP has a significant part of this kind of core practices: small releases, iterations, weeks management, daily meetings, pair programming, TDD, continuous integration, the whole team, informative workspace 
XP advantages
  • XP is a fundamental guidance for managing the rolling wave: fundamental principles, values, and core practices. For some of them, XP is the main reference.
  • Great examples of how principles drive practice selection
  • Offer some process goals descriptions, especially for Inception
  • There are no incompatibilities between XP and any kind of rolling wave approach. XP strongly recommend some core practices but also tell the teams to discover what works for them   
What is missing
  • No explicit process goals except the ones for Inception
  • No real tailoring support, too few references about other options/practices beyond XP
  • Not enough guidance for roadmaps, and no explicit options for more Agile/Lean life-cycles (while implicit support and some useful references exist)
  • Progress toward a Consumable Solution has partial support 


Rolling Wave in DA


·        Information states
  • States are defined primarily at the work level (iterative/agile/lean approach)
  • Life-cycle light milestones ~ real advance of the work states toward the result, a Consumable Solution
  • “Ready” is a clear result of an iterative result of conception, envisioning, looking ahead, just-in-time modeling
  • “Done” is clearly defined as an advance to “consumable solution” with no debts, often feedbacks and validations at all needed levels: life-cycle, iterations, days, day or lower
  •  Integrate often - Continuous integration full cycle 
  • Continuous deployment
  •  A pragmatic and feasible approach
    •  Risks are shifted left
    • Decisions are shifted, not right, but to most appropriate moment, balancing looking ahead with just-in-time
    • Short releases with short and often internal cycles ~ relevant states
    •  Built-in quality
    • Enhanced “Simple Design” - Barely Good Enough - avoids complexity & irrelevance of states
  • State of the product: enhancing the MVP concept & life-cycles evolution in the product life
·        Roadmaps & life-cycles
·        Adapt to context
Value stream support  
  • Work focused on the product as Consumable Solution 
  • Explicit value stream based-approach 
  • Process goals as one of the VS foundations
  • Lightweight milestones 
  • Guideline to streamline the work
 Selecting practices that shape the rolling wave
  • Almost all core practices and reference to many others
  • All XP and Scrum options are referenced or can be used
·        Selecting practices for continuous validation and feedback
  • Almost all core practices and reference to many others
  • All XP and Scrum options are referenced or can be used
·        Collaboration between development teams
  • Guidance for large agile teams (team of teams)
  • Guidance for communities of practices
·        Collaboration with other teams
  • Significant guidance for the collaboration flows with DevOps, IT and Enterprise level teams
DA advantages 
  • in DA you can find almost anything you need to shape your Information Rolling Wave for your custom process, including the ones derived from XP or Scrum.

How to … my custom process


Information states
  • Scrum users – you cannot rely just on iteration & work item states with very vague guidance. 
  • XP users – you have some great guidance, but is incomplete
  • The main recommendation is to use DA guidance to adapt your method and your process to what you need.
  • Life-cycle milestones – use first DA guidance for Agile Basic Life-Cycle (later try to improve to more advanced life-cycles)
  • Ready – use DA guidance about the roadmap trough concept, vision, looking ahead and just-in-time modeling
  • Done – use XP and DA core practices that support Continuous Integration, Continuous deployment 
  • Consumable Solution - use XP and DA guidance for working software and complement with DA guidance about progress toward a Consumable Solution
  • Shift the risks to the left of the life-cycle (milestones, ready, done)
  • Balance looking ahead with just-time for the optimum way to make decisions
  • MVP – use XP and DA guidance
  • Feasibility & pragmatism - use XP and DA guidance for often delivery, built-in quality, and simplicity
  • Build-in quality – if your product has technical debt, the root cause it a deteriorated, messy information inside the product 
  • Value stream support -  see life cycles, Consumable Solution, information states
Roadmaps & life-cycles
Adapt to context
 Value Stream
  • See Information States and Core Practices
  •  
Selecting practices that shape the rolling wave
  • Complement Scrum and XP with other needed practices
  • Do not forget practices for: roadmaps, envisioning, looking ahead, just-in-time reactions
Selecting practices for continuous validation and feedback
  • Complement Scrum and XP with other needed practices
  • Use DA to gradually adopt practices for Continuous Integration, Continuous Deployment
Collaboration between development teams
  • Use DA guidance for large teams/ team of teams and for Communities of Practices
  • Use DA guidance for communities of practices
Collaboration with other teams
  • Use DA guidance for the collaboration flows with DevOps, IT and Enterprise level teams

References (draft) 



Posted by at No comments:
Labels: , , , , , , , , , ,

Saturday, August 18, 2018

Specificity of homo sapiens. Part 2 - Knowledge Work & Software Development


Know thyself - Temple of Apollo, Delphi; Socrate; Aeschylus
Nosce te ipsum! – Cicero
Man, know thyself, and you are going to know the gods - Luxor, External Temple

You may be asking what was the purpose of the previous post "Specificity of homo sapiens. Part 1 - Evolution evidence" in an agile (software development) blog. 
Software development is knowledge work, where the most important factor is the people. Do we, people, use our full potential in this kind of work? We have some industry experience and personal experience but we still struggle with basic things. Where can we find some "hard" evidence about what is important to use full human potential? Evolution. We have the evidence of evolution steps of Homo Sapiens and his predecessors in  thousands of years or more.

I propose a little game - to compare some knowledge work process approaches and what really helped humans to make significant steps forward in their evolution.

NO IMPROVEMENT TEAMS


Process story - I am somehow sad when I saw this kind of phenomenon: a development team with "no improvement", in fact too few improvements in a long period of time (years). There is an individual improvement, there is some adaptation to a specific context, but is no real improvement for the team/organization work approach. When they discover better ways of working, these improvements are not distributed and adopted enough. Yes, they could adopt new and better external tools, but that is like exploiting a natural oasis, not like making an existent hard context livable.   

Evolution - Homo Erectus was "strong and skillful" but <<Not only were they lazy, but they were also very conservative," Shipton said. Their tools stayed the same in both size and composition as the environment around them changed>>  (See Biblio 1)  The problem of H.E. was not really the laziness, but a limited cognitive capability while comparing to us, Sapiens.  

Comments - So, there is a huge waste to have a team that activates in knowledge work area and has too few improvements for long periods of time. That is a waste of talents in Toyota Production System but here is applied not only to professional talents but to most basic human capabilities.
     

SPECIALISTS (ONLY)


Process Story - Domain experience: in knowledge work you need people with cross-functional skills that work collaboratively. There is a pragmatic profile that seems feasible and gets good results: the Generalizing Specialist (explicitly explained by Scott W. Ambler and Disciplined Agile, see Biblio 7). There are still cases where organizations are trying to use specialist-only teams for software development (and silos) and, their results are more than questionable.

Evolution - the "very Sapiens niche" was the one of "generalist specialist", that means to it was able to <<colonized a diversity of challenging environments, including deserts, tropical rainforests, high altitude settings, and the palaeoarctic, but also specialized in its adaptation to some of these extremes.>> (see Biblio 2). 

Comments - when the context is complex: diverse, fluid and extremely adverse and you need to use knowledge to win, you cannot be only "generalist" or only "specialist". You can rely on specialists, but you cannot have only this skills profile or live with the hope that could work well in isolated teams (see below). 

NO COLLABORATIVE WORK   


Process story - There is a correlation between the collaboration level and the result of a team. Teams with rather sporadic collaboration will not use their full potential and their results could be always much better. Just observing the level of direct interactions between team members during a working day you can tell if something is wrong or not.
.
Evolution -  what was different in the case of Homo Sapiens comparing with other human species? Why was able to "conquer the world" and live in all kind of environments even from his early prehistory?  Here are the main factors (see Biblio 2)
  • "extensive cooperation between non-kin individuals"
  • "accumulating, drawing from, and passing down a large pool of cumulative cultural knowledge, in material or idea form, may have been crucial in the creation and maintenance of the generalist-specialist niche by our species in the Pleistocene"  

Comments - what was the fundamental strength of Sapiens from the very beginning ? "extensive collaboration" and management of the knowledge: accumulate, sharing of "a large pool of knowledge". Ignoring collaborative work it is like ignoring that people are Sapiens.
  

ISOLATED TEAMS


Team, team, team .. this is what you heard from most Agile coaches and gurus. Yes, a team of collaborative individuals will perform much better than the same individuals working alone. What about isolated teams? We can have an organization with many teams, but with no interaction between them and not shared knowledge. 
.
Evolution - around 11000 years ago, we, Home Sapiens, have started an accelerated development after 200000 years of very slow evolution. We have done that without a genetic improvement and without any E.T. intervention. At that time people start to live in larger communities where they can better share and accumulate their knowledge (see Biblio 3). It was also easier to value the "specialists", the special talents, that could build communities of practice per "profession". 

Comments - process guidance & methods have highly disregarded the inter-team collaboration and knowledge/experience sharing. It is rather strange and sad: extensive collaboration, communities of practices and inter-team experience sharing was the main factor that starts human civilization (see Biblio 3) after 200000 lost years in isolated small groups and .... we still use the isolated groups approach. 
Isolated teams: that is old prehistory!
There is only one significant exception in process methods for this aspect: Disciplined Agile, where you can find guidance about more levels of collaboration (Delivery, DevOps, IT, Enterprise), guidance about communities of practices, reuse engineering and other aspects (See Biblio 4, Biblio 5, Biblio 6).  


Bibliography and References

[Biblio 1] 
A. Laziness May Have Driven Homo Erectus to Extinction, By Yasemin Saplakoglu, Staff Writer,  August 10, 2018  
B. Acheulean technology and landscape use at Dawadmi, central Arabia, by Ceri Shipton, James Blinkhorn, Paul S. Breeze, Patrick Cuthbertson, Nick Drake, Huw S. Groucutt, Richard P. Jennings, Ash Parton, Eleanor M. L. Scerri, Abdullah Alsharekh, Michael D. Petraglia. PLOS,  Published: July 27, 2018 
[Biblio 2] 
 A. Homo sapiens developed a new ecological niche that separated it from other hominins, July 30, 2018
 B. Defining the ‘generalist specialist’ niche for Pleistocene Homo sapiens, Patrick Roberts & Brian A. Stewart, Nature Human Behaviour; volume 2, pages542–550 (30 July 2018)
[Biblio 3]  - Previous posts related to human evolution and behavior pattern
[Biblio 4]  - The Scope of Disciplined Agile (~ levels of collaboration)
[Biblio 5]  - Disciplined agile - Communities of practice
[Biblio 6]  - Disciplined agile - Reuse Engineering

[Biblio 7]  - Disciplined agile/Agile Modeling - Generalizing Specialist



Posted by at No comments:
Labels: , , , , , , ,
Subscribe to: Posts (Atom)